Spatiotemporally Constrained Action Space Attacks on Deep Reinforcement Learning Agents

TL;DR

This paper introduces novel spatiotemporal attack algorithms on the action space of deep reinforcement learning agents, demonstrating their effectiveness in degrading agent performance and revealing vulnerabilities.

Contribution

It proposes the Myopic and Look-ahead Action Space attack algorithms, incorporating temporal dynamics to enhance attack effectiveness on DRL agents.

Findings

LAS attack significantly outperforms MAS attack with same resources

Attacks exploiting agent dynamics can cause agent failure

Potential use of attacks to identify DRL vulnerabilities

Abstract

Robustness of Deep Reinforcement Learning (DRL) algorithms towards adversarial attacks in real world applications such as those deployed in cyber-physical systems (CPS) are of increasing concern. Numerous studies have investigated the mechanisms of attacks on the RL agent's state space. Nonetheless, attacks on the RL agent's action space (AS) (corresponding to actuators in engineering systems) are equally perverse; such attacks are relatively less studied in the ML literature. In this work, we first frame the problem as an optimization problem of minimizing the cumulative reward of an RL agent with decoupled constraints as the budget of attack. We propose a white-box Myopic Action Space (MAS) attack algorithm that distributes the attacks across the action space dimensions. Next, we reformulate the optimization problem above with the same objective function, but with a temporally coupled constraint on the attack budget to take into account the approximated dynamics of the agent. This leads to the white-box Look-ahead Action Space (LAS) attack algorithm that distributes the attacks across the action and temporal dimensions. Our results shows that using the same amount of resources, the LAS attack deteriorates the agent's performance significantly more than the MAS attack. This reveals the possibility that with limited resource, an adversary can utilize the agent's dynamics to malevolently craft attacks that causes the agent to fail. Additionally, we leverage these attack strategies as a possible tool to gain insights on the potential vulnerabilities of DRL agents.